In this chapter you will work with plant taxonomy data. This type of data does not strictly belong to the field of ecological research, but we hope it provides an easy-to-understand illustration of how to apply the demonstrated methods and how to interpret their results. You will learn how to analyse non-compositional (general) data tables with PCA, how to find which variables separate a priori specified classes (using discriminant analysis), and how to perform unconstrained and constrained ordination with an a priori chosen dissimilarity (distance) measure using, respectively, principal coordinates analysis and distance-based RDA.

Data

The data come from a not-yet-published taxonomic study (Štech et al., unpublished) of several closely related taxa of the genus Melampyrum (hemiparasitic plants in the family Orobanchaceae), where the differences on a wider geographical scale were evaluated using both morphological measurements and molecular data, obtained using the AFLP method (Vos et al. 1995). In this case study, you will be using a subset of 71 individuals originating from 19 populations to address questions relating the variability in morphological characters or in genome to an a priori classification of individuals into four taxa coded as DEG, HOE, NEM, and SUB (this classification was based on other molecular data, namely chloroplast DNA). The questions concerning the definition of well-separated groups from scratch, based on the collected data, are not asked here, but you will be informally evaluating the consistency of a priori defined groups with the variation in the data.